Journal of Emerging Trends in Engineering and Applied Sciences: Research Areas

Article Title: Lung Cancer Detection and Classification based on Image Processing and Statistical Learning
by Md Rashidul Hasan and Muntasir Al Kabir

Abstract:
Lung cancer is one of the death threatening diseases amonghuman beings. Early and accurate detection of lung cancercan increase the survival rate of a lung cancer patient. ComputedTomography (CT) images are commonly used for detectingthe lung cancer. Using a data set of thousands of high-resolution lung scans collected from "Kaggle" competition, we have pre-processed the images by smoothing, enhancement, and segmentation. Then, we measured area, perimeter, skewness, kurtosis, and entropy from the CT images of lung. Finally, we have applied to several statistical method such as logistic regression, Quadratic discriminant analysis, K-nearest neighbors, decision tree, random forest, k-mean clustering, and Support vector machine (SVM) to accurately determine the cancerous cell. These methodshave been tested on 198 slices of CT images of various stages of cancer obtained from "Kaggle" data set and found satisfactory results. Among them SVM identified the cells (whether cancerous or not) with the highest accuracy of 72.2%. In terms of accuracy, random forest (71.2%) and decision tree (71.7%) are also very close to SVM for detecting the cancerous cells.

Keywords: Lung Cancer Detection, Image Processing, and Statistical Learning

Download full paper